Completeness in the Relational Model: a Comprehensive Framework

Completeness is a well known data quality dimension in the area of databases. Intuitively, a database is complete if it represents every fact of the real world coherent with the database semantics, i.e. its intension. In the paper, we provide a comprehensive framework for characterizing completeness in the relational model, investigating several different paradigms typical of database models, such as closed world and open world assumptions, and presence or absence of null values. Furthermore, we introduce an algebra for completeness, in order to address the problem of calculating composition of quality dimensions in queries that include relational operators such as union, difference and cartesian product. Under different assumptions and for the different types of completeness, we provide properties and shortcuts for such an algebra.